---
license: cc-by-nc-4.0
tags:
- emmi-ai
- physics-simulation
- dem
- cfd-dem
- neural-operator
pretty_name: NeuralDEM Dataset
---

# NeuralDEM Dataset

Dataset repository for **NeuralDEM**, containing simulation data for training and evaluating deep learning surrogates for industrial particulate flows and particle-fluid coupled systems.

## Dataset Summary

The NeuralDEM dataset covers two primary physics benchmarks:

1. **Hopper Simulations (Particle Systems)**
   * **Setup**: Hopper domain with a bottom outlet initially loaded with ~250,000 particles discharging over time.
   * **Physics**: Discrete Element Method (DEM) dynamics across diverse hopper geometry angles and particle friction regimes.
2. **Fluidized Bed Reactor (Particle-Fluid Coupled Systems)**
   * **Setup**: Reactor containing ~500,000 particles with uniform fluid (air) injection from the bottom grid.
   * **Physics**: Coupled CFD-DEM multi-physics system over ~160,000 hexahedral CFD grid cells across varying fluid inlet velocities.


For inference scripts, model checkpoints, and simulation rollouts, visit the [NeuralDEM GitHub Repository](https://github.com/Emmi-AI/NeuralDEM).

## License

This dataset is distributed under the [CC-BY-NC 4.0](https://creativecommons.org/licenses/by-nc/4.0/) license.

## Citation instructions
```
@article{alkin2024neuraldem,
  title={{NeuralDEM} for real time simulations of industrial particular flows},
  author={Benedikt Alkin and Tobias Kronlachner and Samuele Papa and Stefan Pirker and Thomas Lichtenegger and Johannes Brandstetter},
  journal={Nature Communications Physics},
  year={2025}
  doi={10.1038/s42005-025-02342-4},
}
```